Message-ID: <CAK7LNASvCha27kU4ipn23uOpNuxkzJrNzWBwYcxN4n=3xtv8SA@mail.gmail.com>
Date:   Wed, 20 Jun 2018 02:34:35 +0900
From:   Masahiro Yamada <yamada.masahiro@...ionext.com>
To:     Paul Burton <paul.burton@...s.com>, Arnd Bergmann <arnd@...db.de>
Cc:     Linux Kbuild mailing list <linux-kbuild@...r.kernel.org>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Linux-MIPS <linux-mips@...ux-mips.org>,
        Ingo Molnar <mingo@...nel.org>,
        Matthew Wilcox <matthew@....cx>,
        Thomas Gleixner <tglx@...utronix.de>,
        Douglas Anderson <dianders@...omium.org>,
        Josh Poimboeuf <jpoimboe@...hat.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Matthias Kaehlcke <mka@...omium.org>,
        He Zhe <zhe.he@...driver.com>,
        Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Michal Marek <michal.lkml@...kovi.net>,
        Khem Raj <raj.khem@...il.com>,
        Christophe Leroy <christophe.leroy@....fr>,
        Al Viro <viro@...iv.linux.org.uk>,
        Stafford Horne <shorne@...il.com>,
        Gideon Israel Dsouza <gidisrael@...il.com>,
        Kees Cook <keescook@...omium.org>,
        Michael Ellerman <mpe@...erman.id.au>,
        Heiko Carstens <heiko.carstens@...ibm.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Paul Mackerras <paulus@...ba.org>,
        linuxppc-dev <linuxppc-dev@...ts.ozlabs.org>
Subject: Re: [PATCH 1/3] kbuild: add macro for controlling warnings to linux/compiler.h

Hi.

2018-06-16 9:53 GMT+09:00 Paul Burton <paul.burton@...s.com>:
> From: Arnd Bergmann <arnd@...db.de>
>
> I have occasionally run into a situation where it would make sense to
> control a compiler warning from a source file rather than doing so from
> a Makefile using the $(cc-disable-warning, ...) or $(cc-option, ...)
> helpers.
>
> The approach here is similar to what glibc uses, using __diag() and
> related macros to encapsulate a _Pragma("GCC diagnostic ...") statement
> that gets turned into the respective "#pragma GCC diagnostic ..." by
> the preprocessor when the macro gets expanded.
>
> Like glibc, I also have an argument to pass the affected compiler
> version, but decided to actually evaluate that one. For now, this
> supports GCC_4_6, GCC_4_7, GCC_4_8, GCC_4_9, GCC_5, GCC_6, GCC_7,
> GCC_8 and GCC_9. Adding support for CLANG_5 and other interesting
> versions is straightforward here. GNU compilers starting with gcc-4.2
> could support it in principle, but "#pragma GCC diagnostic push"
> was only added in gcc-4.6, so it seems simpler to not deal with those
> at all. The same versions show a large number of warnings already,
> so it seems easier to just leave it at that and not do a more
> fine-grained control for them.
>
> The use cases I found so far include:
>
> - turning off the gcc-8 -Wattribute-alias warning inside of the
>   SYSCALL_DEFINEx() macro without having to do it globally.
>
> - Reducing the build time for a simple re-make after a change,
>   once we move the warnings from ./Makefile and
>   ./scripts/Makefile.extrawarn into linux/compiler.h
>
> - More control over the warnings based on other configurations,
>   using preprocessor syntax instead of Makefile syntax. This should make
>   it easier for the average developer to understand and change things.
>
> - Adding an easy way to turn the W=1 option on unconditionally
>   for a subdirectory or a specific file. This has been requested
>   by several developers in the past that want to have their subsystems
>   W=1 clean.
>
> - Integrating clang better into the build systems. Clang supports
>   more warnings than GCC, and we probably want to classify them
>   as default, W=1, W=2 etc, but there are cases in which the
>   warnings should be classified differently due to excessive false
>   positives from one or the other compiler.
>
> - Adding a way to turn the default warnings into errors (e.g. using
>   a new "make E=0" tag) while not also turning the W=1 warnings into
>   errors.
>
> This patch for now just adds the minimal infrastructure in order to
> do the first of the list above. As the #pragma GCC diagnostic
> takes precedence over command line options, the next step would be
> to convert a lot of the individual Makefiles that set nonstandard
> options to use __diag() instead.
>
> [paul.burton@...s.com:
>   - Rebase atop current master.
>   - Add __diag_GCC, or more generally __diag_<compiler>, abstraction to
>     avoid code outside of linux/compiler-gcc.h needing to duplicate
>     knowledge about different GCC versions.
>   - Add a comment argument to __diag_{ignore,warn,error} which isn't
>     used in the expansion of the macros but serves to push people to
>     document the reason for using them - per feedback from Kees Cook.]
>
> Signed-off-by: Arnd Bergmann <arnd@...db.de>
> Signed-off-by: Paul Burton <paul.burton@...s.com>
> Cc: Michal Marek <michal.lkml@...kovi.net>
> Cc: Masahiro Yamada <yamada.masahiro@...ionext.com>
> Cc: Douglas Anderson <dianders@...omium.org>
> Cc: Al Viro <viro@...iv.linux.org.uk>
> Cc: Heiko Carstens <heiko.carstens@...ibm.com>
> Cc: Mauro Carvalho Chehab <mchehab@...nel.org>
> Cc: Matthew Wilcox <matthew@....cx>
> Cc: Matthias Kaehlcke <mka@...omium.org>
> Cc: Arnd Bergmann <arnd@...db.de>
> Cc: Ingo Molnar <mingo@...nel.org>
> Cc: Josh Poimboeuf <jpoimboe@...hat.com>
> Cc: Kees Cook <keescook@...omium.org>
> Cc: Andrew Morton <akpm@...ux-foundation.org>
> Cc: Thomas Gleixner <tglx@...utronix.de>
> Cc: Gideon Israel Dsouza <gidisrael@...il.com>
> Cc: Christophe Leroy <christophe.leroy@....fr>
> Cc: Benjamin Herrenschmidt <benh@...nel.crashing.org>
> Cc: Paul Mackerras <paulus@...ba.org>
> Cc: Michael Ellerman <mpe@...erman.id.au>
> Cc: Stafford Horne <shorne@...il.com>
> Cc: Khem Raj <raj.khem@...il.com>
> Cc: He Zhe <zhe.he@...driver.com>
> Cc: linux-kbuild@...r.kernel.org
> Cc: linux-kernel@...r.kernel.org
> Cc: linux-mips@...ux-mips.org
> Cc: linuxppc-dev@...ts.ozlabs.org
> ---
>
>  include/linux/compiler-gcc.h   | 66 ++++++++++++++++++++++++++++++++++
>  include/linux/compiler_types.h | 18 ++++++++++
>  2 files changed, 84 insertions(+)
>
> diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h
> index f1a7492a5cc8..aba64a2912d8 100644
> --- a/include/linux/compiler-gcc.h
> +++ b/include/linux/compiler-gcc.h
> @@ -347,3 +347,69 @@
>  #if GCC_VERSION >= 50100
>  #define COMPILER_HAS_GENERIC_BUILTIN_OVERFLOW 1
>  #endif
> +
> +/*
> + * turn individual warnings and errors on and off locally, depending
> + * on version.
> + */
> +#define __diag_GCC(version, s) __diag_GCC_ ## version(s)
> +
> +#if GCC_VERSION >= 40600
> +#define __diag_str1(s) #s
> +#define __diag_str(s) __diag_str1(s)
> +#define __diag(s) _Pragma(__diag_str(GCC diagnostic s))
> +
> +/* compilers before gcc-4.6 do not understand "#pragma GCC diagnostic push" */
> +#define __diag_GCC_4_6(s) __diag(s)
> +#else
> +#define __diag(s)
> +#define __diag_GCC_4_6(s)
> +#endif
> +
> +#if GCC_VERSION >= 40700
> +#define __diag_GCC_4_7(s) __diag(s)
> +#else
> +#define __diag_GCC_4_7(s)
> +#endif
> +
> +#if GCC_VERSION >= 40800
> +#define __diag_GCC_4_8(s) __diag(s)
> +#else
> +#define __diag_GCC_4_8(s)
> +#endif
> +
> +#if GCC_VERSION >= 40900
> +#define __diag_GCC_4_9(s) __diag(s)
> +#else
> +#define __diag_GCC_4_9(s)
> +#endif
> +
> +#if GCC_VERSION >= 50000
> +#define __diag_GCC_5(s) __diag(s)
> +#else
> +#define __diag_GCC_5(s)
> +#endif
> +
> +#if GCC_VERSION >= 60000
> +#define __diag_GCC_6(s) __diag(s)
> +#else
> +#define __diag_GCC_6(s)
> +#endif
> +
> +#if GCC_VERSION >= 70000
> +#define __diag_GCC_7(s) __diag(s)
> +#else
> +#define __diag_GCC_7(s)
> +#endif
> +
> +#if GCC_VERSION >= 80000
> +#define __diag_GCC_8(s) __diag(s)
> +#else
> +#define __diag_GCC_8(s)
> +#endif
> +
> +#if GCC_VERSION >= 90000
> +#define __diag_GCC_9(s) __diag(s)
> +#else
> +#define __diag_GCC_9(s)
> +#endif


Hmm, we would have to add this for every release.



> diff --git a/include/linux/compiler_types.h b/include/linux/compiler_types.h
> index 6b79a9bba9a7..313a2ad884e0 100644
> --- a/include/linux/compiler_types.h
> +++ b/include/linux/compiler_types.h
> @@ -271,4 +271,22 @@ struct ftrace_likely_data {
>  # define __native_word(t) (sizeof(t) == sizeof(char) || sizeof(t) == sizeof(short) || sizeof(t) == sizeof(int) || sizeof(t) == sizeof(long))
>  #endif
>
> +#ifndef __diag
> +#define __diag(string)
> +#endif
> +
> +#ifndef __diag_GCC
> +#define __diag_GCC(string)
> +#endif




__diag_GCC() takes two arguments,
so it should be:

#ifndef __diag_GCC
#define __diag_GCC(version, s)
#endif


Otherwise, this would cause warning like this:


arch/arm64/kernel/sys.c:40:1: error: macro "__diag_GCC" passed 2
arguments, but takes just 1
 SYSCALL_DEFINE1(arm64_personality, unsigned int, personality)
 ^~~~~~~~~~







> +#define __diag_push()  __diag(push)
> +#define __diag_pop()   __diag(pop)
> +
> +#define __diag_ignore(compiler, version, option, comment) \
> +       __diag_ ## compiler(version, ignored option)
> +#define __diag_warn(compiler, version, option, comment) \
> +       __diag_ ## compiler(version, warning option)
> +#define __diag_error(compiler, version, option, comment) \
> +       __diag_ ## compiler(version, error   option)
> +


To me, it looks like this is putting GCC/Clang specific things
in the common file, <linux/compiler_types.h> .

All compilers must use "ignored", "warning", "error",
not allowed to use "ignore".



I also wonder if we could avoid proliferating __diag_GCC_*.



>  #endif /* __LINUX_COMPILER_TYPES_H */
> --
> 2.17.1
>


I attached a bit different implementation below.

I used -Wno-pragmas to avoid unknown option warnings.




diff --git a/Makefile b/Makefile
index ca2af1a..d610d81 100644
--- a/Makefile
+++ b/Makefile
@@ -817,6 +817,8 @@ KBUILD_CFLAGS   += $(call cc-option,-Werror=designated-init)
 # change __FILE__ to the relative path from the srctree
 KBUILD_CFLAGS  += $(call cc-option,-fmacro-prefix-map=$(srctree)/=)

+KBUILD_CFLAGS   += $(call cc-option,-Wno-pragmas)
+
 # use the deterministic mode of AR if available
 KBUILD_ARFLAGS := $(call ar-option,D)

diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h
index f1a7492..3f9c1cc 100644
--- a/include/linux/compiler-gcc.h
+++ b/include/linux/compiler-gcc.h
@@ -3,6 +3,8 @@
 #error "Please don't include <linux/compiler-gcc.h> directly, include
<linux/compiler.h> instead."
 #endif

+#include <linux/stringify.h>
+
 /*
  * Common definitions for all gcc versions go here.
  */
@@ -259,6 +261,16 @@
  */
 #define __visible      __attribute__((externally_visible))

+/* turn individual warnings and errors on and off locally */
+#define __diag_gcc(s)  _Pragma(__stringify(GCC diagnostic s))
+
+#define __diag_push()  __diag_gcc(push)
+#define __diag_pop()   __diag_gcc(pop)
+
+#define __diag_ignore(option, comment) __diag_gcc(ignored __stringify(option))
+#define __diag_warn(option, comment)   __diag_gcc(warning __stringify(option))
+#define __diag_error(option, comment)  __diag_gcc(error __stringify(option))
+
 #endif /* GCC_VERSION >= 40600 */
diff --git a/include/linux/compiler_types.h b/include/linux/compiler_types.h
index 6b79a9b..32e354f 100644
--- a/include/linux/compiler_types.h
+++ b/include/linux/compiler_types.h
@@ -271,4 +271,24 @@ struct ftrace_likely_data {
 # define __native_word(t) (sizeof(t) == sizeof(char) || sizeof(t) ==
sizeof(short) || sizeof(t) == sizeof(int) || sizeof(t) ==
sizeof(long))
 #endif

+#ifndef __diag_push
+#define __diag_push()
+#endif
+
+#ifndef __diag_pop
+#define __diag_pop()
+#endif
+
+#ifndef __diag_ignore
+#define __diag_ignore(option, comment)
+#endif
+
+#ifndef __diag_warn
+#define __diag_warn(option, comment)
+#endif
+
+#ifndef __diag_error
+#define __diag_error(option, comment)
+#endif
+
 #endif /* __LINUX_COMPILER_TYPES_H */





Usage is

       __diag_push();
       __diag_ignore(-Wattribute-alias,
                     "Type aliasing is used to sanitize syscall arguments");
              ...
       __diag_pop();




Comments, ideas are appreciated.




-- 
Best Regards
Masahiro Yamada

Powered by blists - more mailing lists